ZATKOFF, District Judge.
Defendant-appellant Akron Firefighters Association, AFL-CIO, Local 330 (“Local 330”) appeals the district court’s order approving a consent decree between plaintiff-appellee United Black Firefighters Association (“UBFA”) and defendant-appellee City of Akron (“City”). The consent decree will increase the percentage of black firefighter/medics who are promoted to the position of lieutenant in the Akron Fire Department by promoting all persons who are ranked numbers one through fifty-nine on the applicable promotions list by August 31,1992, regardless of whether fifty-nine lieutenant positions actually become available during this time.
Local 330 raises two issues for our review: (1) whether the district court’s approval of the consent decree between the City and the UBFA, which requires the City to overfill the position of fire lieutenant and alters future promotion eligibility procedures within the fire department, illegally violates Local 330’s collective bargaining rights and the collective bargaining agreement between the City and Local 330, and (2) whether the overfill provision of the consent decree between the City and the UBFA constitutes a racial preference, and if so, whether the preference violates the Equal Protection Clause of the Fourteenth Amendment? For the reasons stated below, we AFFIRM, in part, and VACATE, in part, the order approving the consent decree and REMAND for further proceedings.
I.
A.
On September 19, 1990, on behalf of all black firefighters who took a promotional examination for lieutenant in 1990, the UBFA and six of its members1 filed a class action against the City, the Akron Civil Service Commission, Akron Fire Chief Ro-manoski, Mayor Donald Plusquellic, and City Personnel Director Richard Pamley. [1003]*1003The UBFA alleged that the promotional test given in 1990 was tainted by racial discrimination and brought actions under Title VII of the Civil Rights Act of 1964, 42 U.S.C. §§ 2000e-2000e-17 and 42 U.S.C. § 1981. The UBFA also brought an action under 42 U.S.C. § 1983 for alleged denial of due process and an action under a contract theory for breach of a 1986 consent decree concerning the promotional process of firefighters to the position of lieutenant. In connection with the 1986 consent decree, the UBFA specifically alleged that defendants violated the consent decree by failing “not to discriminate [in promotions] and to set up a nondiscriminatory process for promotions.” J.A. 17. On December 10, 1990, the UBFA amended its complaint to include Local 330 as a defendant. Local 330 is the authorized and exclusive representative of Akron firefighters.
On February 12, 1991, the UBFA and the City filed a proposed consent decree with the district court. Local 330 was not a signatory to the consent decree. On March 6, 1991, the district court, pursuant to Federal Rule of Civil Procedure 23, certified the class of plaintiffs.2
A fairness hearing was scheduled for March 21, 1991. Notice of the consent decree and fairness hearing was given by publication, by individual mailings to black firefighters, and by posting the decree at each fire station. On March 24, 1991, Local 330 filed objections to the proposed consent decree. The fairness hearing was conducted on March 21, 1991, and continued on April 8, 1991, with all interested persons being given the opportunity to present evidence. The district court approved the consent decree and ordered it filed with the court on April 22, 1991. This timely appeal followed.
B.
In April of 1991, the Akron Fire Department was comprised of 381 firefighters and medics, 89 of whom were black and 292 of whom were white. There were a total of 87 fire lieutenants, 6 of whom were black and 81 of whom were white.
In the summer of 1990, the City conducted the promotional examination at issue. The 1990 Fire Lieutenant’s Examination consisted of four components: (1) service ratings, (2) the Assessment Center, (3) a written examination, and (4) seniority points. An applicant’s service rating is determined by six-month evaluations.
Persons who passed the examination were ranked according to their scores on the examination and placed on a promotions list effective August 31, 1990. Persons on the list are to be promoted to lieutenant in rank order when positions become available. Thus, the promotional list not only defines those qualified for promotion, but also governs the order in which candidates are promoted. The promotional list at issue has a life of two years, after which all persons remaining on the list must take another examination to determine their eligibility to be promoted and, if eligible, in what order they will be promoted.
A total of 137 applicants, 38 black and 99 white, took the 1990 examination. 117 persons passed the examination. 76.3 percent of the black applicants or 29 of the 38 black applicants passed, and 89.3 percent of the white applicants or 89 of the 99 white applicants passed. Black applicants ranked in the top 50 positions at positions 21, 25, 41, 43, 47 and 50. Fire Chief Romanoski projected a future need to fill 47 to 49 lieutenant positions over the two-year span of the promotions list. However, the consent decree provides that
[1004]*1004the City of Akron agrees to promote individuals in consecutive rank as they appear on the existing eligibility list to the position of Fire Lieutenant as normal vacancies occur; however, the City shall promote no less than the number of individuals necessary to reach the fifty-ninth person on the list no later than August 31, 1992, even if Lieutenants positions must be overfilled. The list of Fire Lieutenant will expire on August 31, 1992, with at least fifty-nine appointments in rank order made.
J.A. 27-28. Consequently, under the consent decree, the City is required to exceed its need for lieutenants; that is, the City will overfill its lieutenant ranks by a margin of 10 to 12.
If the City promotes only those candidates on the list ranked 1 through 49, in accordance with the most liberal allowance predicted by Chief Romanoski, 5 black candidates will be promoted. Under the same scenario, 42 white candidates will be promoted. However, under the consent decree, by promoting persons ranked 1 through 59, nine black candidates will be promoted while 50 white candidates will be promoted. This result would be accomplished because the persons ranked 50, 54, 58, and 59 are black while the persons ranked 51 through 53 and 55 through 57 are white.
The consent decree also eliminates the “service rating” component of all promotional examinations administered on or before August 31, 1994. In addition, the consent decree establishes a procedure for choosing an expert to develop the “Assessment Center” exercise for future promotional examinations. The Assessment Center component is a more subjective part of the examination and, in part, requires the candidate to give an oral presentation. The procedure requires the City and the UBFA to select an expert who will devise the Assessment Center. If they cannot agree on an expert, the district court will choose one. Local 330 may participate as an equal party in the selection of the expert, but to do so it must waive its right to challenge any promotional examination later administered pursuant to the selected expert’s proposal.
II.
A consent decree is a final judicial order. Williams v. Vukovich, 720 F.2d 909, 920 (6th Cir.1983). Thus, this court has jurisdiction to review a district court’s order approving a consent decree. Id. A consent decree is also a compromise among parties after each has assessed the risks and expenses of further litigation. Id. at 922.
When reviewing this consent decree, this court must consider that under Fed. R.Civ.P. 23(e), a class action may not be compromised without the approval of the district court. Bailey v. Great Lakes Canning, Inc., 908 F.2d 38, 42 (6th Cir.1990). The district court may not approve the consent decree unless it is fair and reasonable to those it affects and is consistent with the public interest. Id. at 42; Williams, 720 F.2d at 921. Moreover, judicial approval may not be obtained for a consent decree that is illegal or is the product of collusion. Williams, 720 F.2d at 920.3 When reviewing a district court’s order approving a consent decree in light of these considerations, this court will not set aside a consent decree unless the district court has abused its discretion. Bailey, 908 F.2d at 42. However, “the validity of a consent decree’s affirmative action plan is a question of law” and will be reviewed de novo. Donaghy v. City of Omaha, 933 F.2d 1448, 1458 (8th Cir.1991).
Before addressing the merits of this case, we must examine the UBFA’s argument that a former adjudication precludes Local 330 from raising its constitutional and statutory objections to the consent de[1005]*1005cree. Specifically, the UBFA argues that the issues in this case were litigated and decided in its favor when the 1986 consent decree was proposed and that the 1986 consent decree is fully incorporated into the 1991 consent decree at issue in this case.
In 1986, Local 330 objected to the then proposed consent decree on the basis that it violated the existing collective bargaining agreement. Local 330 presently objects to the 1991 consent decree on the basis that it violates Ohio law as it applies to the separate and different collective bargaining agreement now in effect. Consequently, neither res judicata nor collateral estoppel applies. See NAACP, Detroit Branch v. Detroit Police Officer Association, 821 F.2d 328, 330 (6th Cir.1987) (before collateral estoppel can apply, the precise issue raised in the present case must have been raised in a prior proceeding); Brown v. Felsen, 442 U.S. 127, 131, 99 S.Ct. 2205, 2209, 60 L.Ed.2d 767 (1979) (“under res judicata, a final judgment on the merits bars further claims by parties or their privies based on the same cause of action.”) (emphasis added).
Moreover, the 1986 consent decree was concerned with remedying alleged racial discrimination prior to 1986. The present consent decree is designed to remedy alleged racial discrimination in connection with a 1990 promotional examination and which occurred after the 1986 consent decree became effective. Thus, as earlier stated, neither res judicata nor collateral estoppel applies. Id.
The UBFA and the City both join in a second procedural argument — that Local 330 failed to raise below its objection that matters affecting promotions are mandatory subjects of bargaining and, therefore, are not properly the subject of a consent decree between the City and the UBFA. On the contrary, Local 330 did state in its objections to the consent decree and in its accompanying brief that the consent decree impermissibly modifies the bargaining rights of Akron Firefighters without the participation of Local 330, the exclusive bargaining agent for Akron firefighters. Although this statement is not specific, it is sufficient to preserve the issue for appeal.
C.
Local 330’s first argument on the merits is that the district court erred in holding that the consent decree does not illegally impinge upon its role as exclusive collective bargaining representative. Local 330 argues that the consent decree violates Ohio Rev.Code § 4117.08(A) and thus is invalid. Ohio Rev.Code § 4117.08(A) provides:
All matters pertaining to wages, hours, or terms and other conditions of employment ... are subject to collective bargaining between the public employer and the exclusive representative, except as otherwise specified in this section.
Relying on DeVennish v. Columbus, 57 Ohio St.3d 163, 566 N.E.2d 668 (1991), Local 330 argues that promotions are a mandatory subject of bargaining; therefore, it is argued, the consent decree, which establishes specific procedures for promotions to lieutenant effective until 1994, usurps Local 330’s role as the exclusive bargaining representative of Akron firefighters. In addition, Local 330 argues that in executing its duties under the consent decree, the City commits an unfair labor practice in violation of Ohio Rev.Code § 4117.11(A)(5) by failing to bargain in good faith with Local 330 on the issue of promotions.
In their rebuttal to Local 330’s argument, both the City and the UBFA argue that promotions are simply a permissible, not mandatory, subject of bargaining and that because the present collective bargaining agreement does not address promotions, the consent decree is not illegal. The City and the UBFA rely, in part, on Ohio Rev.Code' § 4117.08(C)(5) which states:
Unless a public employer agrees otherwise in a collective bargaining agreement, nothing in Chapter 4117. of the Revised Code impairs the right and responsibility of each employer to: (5) Suspend, discipline, demote, or discharge for just cause, or lay off, transfer, assign, [1006]*1006schedule, promote, or retain employees _
(Emphasis added). Appellees also argue that DeVennish v. Columbus, supra, supports their position that promotions are a permissible, not mandatory, subject of bargaining.
In City of Cincinnati v. Ohio Council 8, 61 Ohio St.3d 658, 576 N.E.2d 745 (1991), the Ohio Supreme Court explicitly states in a footnote:
These (the subjects enumerated in Ohio Rev. Code § 4117.08(C)) are the permissive subjects for bargaining designated by statutes. It is significant that, while the statute establishes that there is no duty to bargain over the enumerated subjects, it does not prohibit such bargaining. Similarly, there should be no bar to bargaining over permissive subjects not enumerated in this statute.
Id. 576 N.E.2d at 752, n. 2.
Clearly, the subject of promotions, which is enumerated in Ohio Rev.Code § 4117.08(C), is simply a permissible and not mandatory subject of bargaining. Therefore, the City is not required to bargain with Local 330 over this issue.
D.
The second issue before us is whether the consent decree’s plan to promote up to the rank of 59 on the promotions list regardless of the number of positions which become available by August 31, 1992, violates the Equal Protection Clause of the Fourteenth Amendment by showing an impermissible preference to black firefighters. In relevant part, the Fourteenth Amendment provides:
No State shall ... deny to any person within its jurisdiction the equal protection of the laws.
Local 330 correctly asserts that under City of Richmond v. Croson, 488 U.S. 469, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989), the Supreme Court interpreted the Equal Protection Clause to provide that any racial classification or preference, even a benign racial classification or preference otherwise known as an affirmative action plan, of a state actor is' subject to strict scrutiny.4 Croson reaffirmed the plurality opinion in Wygant v. Jackson Board of Education, 476 U.S. 267, 106 S.Ct. 1842, 90 L.Ed.2d 260 (1986), in which four justices applied strict scrutiny to an affirmative action plan.
For a state’s racial classification or preference to withstand constitutional scrutiny under Croson (1) the state actor must have a compelling governmental interest, and (2) the racial preference must be narrowly tailored to achieve that compelling interest. Croson, 488 U.S. at 505-07, 109 S.Ct. at 727-28; Long v. City of Saginaw, 911 F.2d 1192, 1195 (6th Cir.1990). Under the second prong of the Croson test, a classification or preference is narrowly tailored if the classification or preference remedies past racial discrimination against that race or races of persons that the classification or preference favors. Croson, 488 U.S. at 507, 109 S.Ct. at 728. Societal discrimination alone is not enough. That group whom the classification or preference favors must have suffered discrimination within the very governmental division that has instituted the affirmative action plan. Id. at 507, 109 S.Ct. at 728; Long, 911 F.2d at 1196.
Local 330 argues that the consent decree’s provision mandating promotions to rank 59 contains a racial preference unsupported by evidence of past discrimination with the Akron Fire Department. On the other hand, the UBFA argues that the issue of racial discrimination was decided [1007]*1007under the 1986 consent decree in favor of the UBFA, and, therefore, Local 330 is collaterally estopped from asserting this objection. However, as previously discussed, the issue in this case is discrimination in connection with the 1990 examination, not alleged discrimination leading up to the 1986 consent decree. Therefore, the UBFA’s argument is meritless.
1.
In order to correctly analyze Local 330’s Equal Protection challenge to the consent decree, we must first determine whether the consent decree embodies a racial classification or preference. Local 330 makes the following argument regarding its assertion that the consent decree contains a racial preference: The consent decree favors black candidates because the decision to promote persons ranked 1 through 59 on the list and not through persons ranked 60 or beyond on the list is because the candidate at the 59th position is black while the candidate at the 60th position is white. As a result, the white candidate ranked 60 will not be promoted while the black candidate ranked 59 will be promoted. The consequence of overfilling the lieutenant positions will be fewer positions available for those candidates who take the next Fire Lieutenant’s Examination.
In contrast, the City argues that the consent decree institutes no racial classification or preference but simply expedites the promotions process. The City notes that both black and white candidates alike will be promoted from the ranks of 50 through 59, and all candidates regardless of race seeking to take the 1992 Fire Lieutenant’s Examination will have to compete for fewer positions due to the overfill from the 1990 examination. Thus, the City argues, the consent decree embodies no affirmative action plan, so strict scrutiny does not apply. Moreover, the City strenuously denies any past wrongdoing in the form of racial discrimination and adds the disingenuous argument that the purpose of the consent decree is exclusively to avoid bitter, expensive, protracted litigation;5 therefore, the City argues, the consent decree is unrelated to any attempt to increase the number of black lieutenants in the Akron Fire Department.
It is difficult to characterize this case as one involving a racial classification; it is not so difficult, however, to characterize this case as one involving a racial preference. Candidates who rank in positions 49 through 59 are both black and white; thus, this group does not constitute a suspect class. Nevertheless, both appellees ignore, but we cannot, the manner in which it was decided to promote to the rank of 59. The record clearly reveals that the decision to promote 59 firefighters and no more was based on the fact that the firefighter ranked 60 is white, and that by promoting 59 firefighters, the Akron Fire Department can increase the percentage of successful black applicants promoted to lieutenant from 13.8 percent to 31 percent. The black firefighter ranked 59 was shown a racial preference in the form of a guaranteed promotion while the qualified white firefighter ranked 60 was not.6 This is a racial preference. See Croson, 488 U.S. at 494, 109 S.Ct. at 721 (“ ‘[t]he guarantee of equal protection cannot mean one thing when applied to one individual and something else when applied to a person of another col- or.’ ”) (quoting University of California Regents v. Bakke, 438 U.S. 265, 289-290, [1008]*100898 S.Ct. 2733, 2748, 57 L.Ed.2d 750 (1978)). Id.
As for the City’s argument that it entered the consent decree strictly for practical purposes, and thus the consent decree was not directed at increasing the number of black lieutenants by promoting up to the 59th position on the list, the City simply ignores the fact that the express purpose of the UBFA’s action was to address alleged racial discrimination in promotions to fire lieutenant and that the consent decree embodies a settlement regarding this underlying claim of its action. Although the City may disclaim in its consent decree any past discrimination for purposes of liability, such a disclaimer does not preclude the court from considering the issues involved in the underlying claim when reviewing the validity of the consent decree. See generally, Stotts v. Memphis Fire Dept., 679 F.2d 541, 553 n. 10 (6th Cir.1982), rev’d on other grounds; 467 U.S. 561, 104 S.Ct. 2576, 81 L.Ed.2d 483 (1984); Stuart v. Roache, 951 F.2d 446, 453 (1991); Donaghy v. City of Omaha, 933 F.2d 1448, 1460 (8th Cir.1991).
What is more, appellees specifically admitted at oral argument that racial considerations dominated negotiations over the consent decree.
Regardless of the fact that positions 59 and 60 are occupied by individuals of different races, the consent decree still embodies a racial preference. The City’s clear purpose for overfilling is the attainment of a higher percentage of black lieutenants. Therefore, irrespective of the races of persons 59 and 60, the City has clearly manifested a preference for black firefighters over white firefighters in this promotion scheme.
The City’s argument that because the consent decree benefits both blacks and whites it does not represent a racial preference lacks merit. The near identical argument was proffered in Loving v. Virginia and was unanimously rejected by the Supreme Court. 388 U.S. 1, 87 S.Ct. 1817, 18 L.Ed.2d 1010 (1967) (striking anti-miscegenation statute, though state argued that it affected equally the marital rights of both black and white persons). The City cannot disguise or conceal its racial preference by also overfilling with white applicants.
2.
Having determined that the consent decree does in fact embody a racial preference, we turn to the issue of whether the racial preference must comport with strict scrutiny and whether it, in fact, does. The UBFA argues that a voluntary consent decree is not subject to strict scrutiny. Similarly, the district court held:
A voluntary agreement, as herein submitted by the plaintiff class and the City of Akron, need not be based on evidence to support a prima facie case or even an “arguable violation” on the part of the city with respect to the allegedly racially discriminatory nature of the competitive process for promotion to the rank of Lieutenant in the Fire Department.
J.A. 33-34.
The UBFA argues, and the district court held, that to be judicially approved, a voluntary consent decree need only be “fair, adequate, reasonable and consistent with public policy.” Essentially, the district court and the UBFA take the same position that a voluntary consent decree that embodies a state actor’s racial preference need only comport with the general standards of review for a consent decree and is exempt from those requirements set forth in Croson. However, Croson holds that all racial preferences instituted by a state actor, even those designed to achieve “racial balance” in a state actor’s workforce, are subject to strict scrutiny, which requires, in part, evidence of past discrimination. Croson simply does not exempt consent decrees from its requirements. See, e.g., Stuart, 951 F.2d at 449; Davis v. City of San Francisco, 890 F.2d 1438, 1445 (9th Cir.1989).7
[1009]*1009As previously mentioned, the Supreme Court clearly held in Croson that a state actor’s racial preference survives strict scrutiny when (1) a compelling governmental interest underlies its racial preference, and (2) the preference is narrowly tailored to achieve that interest. Croson, 488 U.S. at 505-07, 109 S.Ct. at 727-28. Under the first prong of the Croson test, a state actor possesses a compelling state interest when its concern is with remedying past discrimination. However, there is no clear majority mandate in Croson on the degree or quantum of evidence required to identify actual past discrimination.8
In Croson, the plaintiffs asserted an Equal Protection challenge to a city ordinance requiring prime contractors to whom the City awarded construction contracts to subcontract at least 30 percent of the dollar amount of each City contract to businesses owned by minorities. In striking down the ordinance, the Court held that there was an insufficient showing of past discrimination against minority subcontractors within the city of Richmond to justify the racial preference. Croson, 488 U.S. at 502-03, 109 S.Ct. at 725.
Quoting from Wygant v. Jackson Board of Education, 476 U.S. at 277, 106 S.Ct. at 1848, the Court in Croson focused on the lack of a “strong basis in evidence for the conclusion that remedial action was necessary.” Id. 488 U.S. at 500, 109 S.Ct. at 724 (Wygant, 476 U.S. at 277, 106 S.Ct. at 1848 (plurality opinion)). The Court further stated, “[tjhere is nothing approaching a prima facie case of a constitutional or statutory violation by anyone in the Richmond construction industry.” Croson, 488 U.S. at 500, 109 S.Ct. at 724 (citing Wygant, 476 U.S. at 274-75, 106 S.Ct. at 1847) (emphasis in original).
Similarly, in Long v. City of Saginaw, 911 F.2d 1192, 1196 (6th Cir.1990), we used the same “strong basis in evidence” language from Wygant to determine whether a compelling governmental interest existed for an affirmative action program designed to increase the number of black officers in the Saginaw Police Department. We stated that
[i]n considering the presence or absence of a compelling governmental interest necessary for justifying affirmative action remedial programs incorporating racial classifications, the teachings of Wy-gant are directory:
“[I]n such a ease, the trial court must make a factual determination that the employer had a strong basis in evidence for its conclusions that remedial action was necessary....”
Id. at 1196 (quoting Wygant, 476 U.S. at 277, 106 S.Ct. at 1848). In addition we also stated that “the Saginaw Police Department must ensure, before it embarks upon an affirmative action program, that it has convincing proof of prior discrimina-tion_” Long, 911 F.2d at 1197 (emphasis added).
Additionally, other circuits, after reviewing Wygant and Croson, have also focused and relied on the “strong basis in evidence” language to determine if sufficient evidence of past discrimination existed to justify a racial preference. Stuart, 951 F.2d at 449; Peightal v. Metropolitan Dade County, 940 F.2d 1394, 1401 (11th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 969, 117 L.Ed.2d 134 (1992); Donaghy v. City of Omaha, 933 F.2d 1448, 1458 (8th Cir.1991); Davis v. City and County of San Francisco, 890 F.2d 1438, 1447 (9th Cir.1989), cert. denied, — U.S. -, 111 S.Ct. 248, 112 L.Ed.2d 206 (1990). Accordingly, based on the explicit language in [1010]*1010Croson, Wygant, and Long, and on the great weight of authority from other circuits, we conclude that the “strong basis in evidence” test, also referred to by this Court as the “convincing evidence” test, is the proper measure for determining whether a state actor has adequately demonstrated Croson’s past intentional discrimination requirement.
In addition to determining the quantum of evidence necessary, we must also decide the methods by which the state actor may comply with the “strong basis in evidence” standard. Croson indicates that statistical methods, if used and analyzed properly, can be employed to identify past discrimination under the “strong basis in evidence” standard:
In the employment context, we have recognized that for certain entry level positions or positions requiring minimal training, statistical comparisons of the racial composition of an employer’s work force to the racial composition of the relevant population may be probative of a pattern of discrimination. But where special qualifications are necessary, the relevant statistical pool for purposes of demonstrating discriminatory exclusion must be the number of minorities qualified to undertake the particular task.
Croson, 488 U.S. at 501-02, 109 S.Ct. at 725 (O’Connor, J., majority opinion) (citations omitted).
In analyzing the statistical data presented in Croson, the Court held that the comparison between the percentage of minorities awarded a contract to the percentage of minorities in the Richmond community had little probative value in determining whether minority subcontractors had experienced discrimination in Richmond because “the City does not know how many [minority business enterprises] in the relevant market are qualified to undertake prime or subcontracting work in public construction projects.” Id. at 502, 109 S.Ct. at 725. In declining to accept the statistical disparity offered in Croson as sufficient to support the affirmative action plan, the Court implied that where a gross statistical disparity arises from a comparison of the percentage of minorities presently employed to the percentage of qualified minorities in the relevant workforce, sufficient evidence may exist to justify the affirmative action plan. Id.
Similarly, in Long, we also considered statistical data and found that the statistical evidence failed to demonstrate past discrimination in the Saginaw Police Department. 911 F.2d at 1199. We reached this conclusion based on our finding that the City failed to establish a prima facie case of intentional discrimination because, as in Croson, the City failed to use the relevant statistical pool with which to make comparisons. Id. In this regard, we stated that the City of Saginaw unsuccessfully “advanced an undifferentiated workforce to support a prima facie case of employment discrimination.” Id. However, we also stated that “a statistical effort ..., if valid, could afford the required probative evidence in the employment policies of the City’s police department.” Id. at 1198. Other circuits have also determined that proper statistical evidence may establish a prima facie case of employment discrimination and may be sufficient to show a “strong basis in evidence” for concluding that remedial action is constitutionally justified. See Stuart, 951 F.2d at 452; Peightal v. Metropolitan Dade County, 940 F.2d at 1401; Davis, 890 F.2d at 1447.
In the present case, although the district court erroneously stated that the consent decree could be approved without evidence of past discrimination, the district court nevertheless engaged in an analysis of some of the statistical data in this case. Specifically, the district court focused on the fact that
if one were to assume that the number to be promoted from the list would stop at forty-six, only four out of a total of twenty-nine black applicants would receive promotions or 13.8 percent of the black applicants that successfully passed the examination. Under the same example, forty-two out of the eighty-nine white applicants who passed the test would be promoted, which causes the appointment [1011]*1011of 47.2 percent of the successful white applicants.
J.A. 36. The district court concluded that given the disparity between the ratio of blacks to whites being promoted viewed from the standpoint of those who received a passing grade and were promotable, the Court finds that there was a reasonable basis for the City of Akron defendants to enter into a consent decree.
J.A. 37.
The statistical analysis that the district court used in the present case deviates from the method generally considered probative of intentional discrimination. The method generally used is to compare the minority percentage in the relevant statistical pool to the minority percentages in the group of persons selected for the positions at issue. See Croson, 488 U.S. at 501-03, 109 S.Ct. at 725-26; Hazelwood School District v. United States, 433 U.S. 299, 308, 97 S.Ct. 2736, 2742, 53 L.Ed.2d 768 (1977); Long, 911 F.2d at 1199. The relevant statistical pool is comprised of all persons qualified for the position at issue. Roughly the same percentage of minorities observed in the relevant statistical pool should also be observed in the group of persons selected for the position at issue. Id. Where a gross disparity exists between the expected percentage of minorities selected and the actual percentage of minorities selected, then prima facie proof exists to demonstrate intentional discrimination in the selection of minorities to those particular positions. See Croson, 488 U.S. at 501, 109 S.Ct. at 725; Hazelwood, 433 U.S. at 307-309, 311 n. 17, 97 S.Ct. at 2741-42, 2743 n. 17 (two to three “standard deviations” from the expected percentage of minorities in the workplace indicates a significant disparity); Long, 911 F.2d at 1199. Such prima facie proof presents a strong basis in evidence to support a finding of a compelling governmental interest. See Croson, 488 U.S. at 503, 109 S.Ct. at 726; Long, 911 F.2d at 1199.
Of course, a gross disparity is not conclusive as to a finding of discrimination, because other factors unrelated to race may account for the disparity. Long, 911 F.2d at 1198. Therefore, those opposing the affirmative action plan may present evidence to rebut the inference, thus defeating the validity of the affirmative action plan. See Stuart, 951 F.2d at 453.
E.
In summary, with respect to Local 330’s claims involving its collective bargaining agreement, we conclude that the district court did not err in finding that the consent decree does not impermissibly infringe upon Local 330’s role as exclusive bargaining agent. Thus, we affirm the district court’s order approving the consent decree as it relates to the elimination of the service rating component of promotional exams and the method for choosing an Assessment Center Exercise.
With respect to Local 330’s constitutional argument, we conclude that the district court erred in entering the consent decree in the absence of proof of past discrimination and that there is not enough evidence in the record to determine if the affirmative action plan comports with strict scrutiny, i.e., whether it is narrowly tailored to achieve a compelling governmental interest. In this respect, we note that the record does not provide enough information for us to analyze statistical data in this case in an adequate manner.
To evaluate adequately any statistical data, there must be (1) evidence identifying the basic qualifications of a fire lieutenant and (2) a determination, based on these qualifications, of the relevant statistical pool with which to make the appropriate comparisons. See Croson, 488 U.S. at 501-02, 109 S.Ct. at 725; Long, 911 F.2d at 1199. In addition, if a non-traditional method of statistical analysis is used, there must be evidence to support the validity of the method in order for this court to review it. It is important to observe that statistical data is not the only means by which discrimination can be shown. See, e.g., Black Law Enforcement Officers Assoc. v. City of Akron, Nos. C84-2974, C85-2781, slip op., 1989 WL 168010 (N.D.Ohio August [1012]*10129, 1989) (unpub.) (witnesses testified that the police department would allow the promotional list to expire rather than promote the next person on the list who was black), aff'd, 920 F.2d 932 (Table) (6th Cir.1990).9 In this case, the discrimination necessary to justify the affirmative action plan must have occurred after the 1986 consent decree and must have tainted the promotional process for the position of fire lieutenant.
III.
For all of the foregoing reasons, the district court’s order approving the consent decree between the United Black Firefighters Association and the City of Akron is hereby AFFIRMED, in part, and VACATED, in part, and this case is REMANDED for further proceedings consistent with this opinion.